Telling Interviewers About Sexual Abuse

CV180
Telling Interviewers About Sexual Abuse
Child Maltreatment
Volume 14 Number 1
February 2009 100-113
© 2009 Sage Publications
10.1177/1077559508318398
http://cm.sagepub.com
hosted at
http://online.sagepub.com
Predictors of Child Disclosure at Forensic Interviews
Tonya Lippert
Dallas Children’s Advocacy Center
Theodore P. Cross
RTI International
Lisa Jones
Wendy Walsh
University of New Hampshire
This study aims to identify characteristics that predict full disclosure by victims of sexual abuse during a forensic interview.
Data came from agency files for 987 cases of sexual abuse between December 2001 and December 2003 from Children’s
Advocacy Centers (CACs) and comparison communities within four U.S. states. Cases of children fully disclosing abuse
when interviewed were compared to cases of children believed to be victims who gave no or partial disclosures. The
likelihood of disclosure increased when victims were girls, a primary caregiver was supportive, and a child’s disclosure
instigated the investigation. The likelihood of disclosure was higher for children who were older at abuse onset and at
forensic interview (each age variable having an independent effect). Communities differed on disclosure rate, with no
difference associated with having a CAC. Findings suggest factors deserving consideration prior to a forensic interview,
including organizational and community factors affecting disclosure rates.
Keywords:
sexual abuse; child disclosure; forensic interview; Children’s Advocacy Center
“Besides, no one ever keeps a secret so well as a
child.”
Victor Hugo, Les Miserables
V
ictor Hugo’s insight succinctly captures the challenge of identifying child sexual abuse: victims will
often keep it a secret for a long time or forever. Research
on children and adults indicates that children often significantly delay disclosure of sexual abuse or keep the
abuse a secret into adulthood (Finkelhor, Hotaling,
Lewis, & Smith, 1990; Smith et al., 2000; Sorenson &
Snow, 1991). Yet child disclosure is the single most significant means by which sexual abuse is discovered (see
Goodman-Brown, Edelstein, Goodman, Jones, & Gordon,
2003), and disclosure at a forensic interview is often
critical to police and child protective services’ (CPS)
investigation of the abuse (see, e.g., Lawson & Chaffin,
1992; Pence & Wilson, 1994).
100
This article focuses on understanding what might
facilitate children’s disclosure at a forensic interview. It
examines several child, abuse, family, and suspect characteristics, replicating previous studies. Adding to previous research, it also explores the unique effects of
children’s age both at the forensic interview and at the
onset of the abuse. These unique effects are helpful to
Authors’ Note: For the purposes of compliance with Section 507 of PL
104-208 (the “Stevens Amendment”), readers are advised that 100% of
the funds for this program are derived from federal sources (this project
was supported by Grants No. 1999-JP-FX-1101, 01-JN-FX-0009,
2002-J W-BX-0002 awarded by the Office of Juvenile Justice and
Delinquency Prevention, Office of Justice Programs, U.S. Department
of Justice). The total amount of federal funding involved is $1,923,276.
Points of view or opinions in this document are those of the author and
do not necessarily represent the official position or policies of the U.S.
Department of Justice. Correspondence concerning this article should
be addressed to Tonya Lippert, 2800 N. Vancouver Ave., Suite 201,
Portland, OR 97227; e-mail: [email protected]
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 101
understand, as they suggest that a child’s disclosure of
abuse can reflect both a child’s present developmental
level and the child’s developmental level at the time his
or her abuse started. This article also analyzes how a preinvestigation disclosure is related to disclosure at a
forensic interview and explores how the use of one particular practice innovation, Children’s Advocacy Centers
(CACs), may affect disclosure. Finally, it examines how
disclosure rates can vary across different communities.
Understanding which child victims are more prone to
disclose during a forensic interview could help shed light
on the nature of disclosure, inform investigators’ decision making, and suggest new practice methods to facilitate disclosure.
Children’s disclosure of sexual abuse can occur various ways. Children may first make a full or partial disclosure to a family member or another individual (e.g.,
teacher, counselor) (Sgroi, Blick, & Porter, 1982).
Alternatively, physical signs, child or suspect behavior,
or other factors may lead to concerns that a child may
have been sexually assaulted (see, e.g., BrilleslijperKater, Friedrich, & Corwin, 2004; Pipe et al., 2007;
Sorenson & Snow, 1991). The initial disclosure or other
evidence of abuse usually leads to someone’s reporting
their concerns, which can then lead to a formal investigation (see, e.g., Finkelhor, Cross, & Cantor, 2005, 2006).
When child sexual abuse is investigated, children who
are old enough to disclose are most often interviewed by
a police officer, a CPS investigator, a child interview specialist at a CAC, or a member of a multidisciplinary team
at a hospital clinic (see, e.g., Pence & Wilson, 1994).
Children respond various ways to the interview—fully or
partially disclosing abuse, giving no disclosure of abuse
(e.g., declining to answer questions on the subject), or
denying any abuse or recanting a previous disclosure
(Sorenson & Snow, 1991).
Whether verbally expressive children will disclose
sexual abuse at a forensic interview could depend on
many factors, including their understanding and memory
of the abuse, their emotional reactions to it, their concerns about the consequences of both disclosure and
nondisclosure, and the extent to which supportive family
members and effective child abuse investigators and
evaluators are involved (see, e.g., Hershkowitz, Horowitz,
& Lamb, 2005, 2007; London, Bruck, Ceci, & Shuman,
2005, 2007: Pipe et al., 2007; Saywitz, Esplin, &
Romanoff, 2007). Children could be more apt to disclose
at the forensic interview if they had disclosed to someone else beforehand (see, e.g., London et al., 2007). The
disclosure rate may also be influenced by a range of
different community or agency variables, including the
characteristics of cases reported to officials within the
community, the type of case typically referred for a forensic interview, the criteria the community uses to determine whether abuse occurred, and the skill of the forensic
interviewers. One question concerns whether the use of
CACs, a major paradigm change designed to facilitate
child abuse investigation, may influence child disclosure.
Previous Research on
Predictors of Child Disclosure
Previous research has explored how child, family, and
investigation variables are related to disclosure, but
many questions about what explains disclosure remain.
Studies have examined how child age, child gender, suspect relationship to child, abuse severity, caregiver support, and initial disclosure relate to child disclosure at a
forensic interview. Other studies have examined predictors of pre-investigation disclosures, which is meaningful but less relevant to the current study. Pipe et al.
(2007) and Sorenson and Snow (1991) studied cases at
CACs but made no comparison between CACs and other
agencies on disclosure rates. We are aware of no studies
that examined suspect age as a predictor of child disclosure, nor any study that compares disclosure rates across
different agencies and communities. Note that comparing
results across studies of different communities would be
difficult because of variation with respect to sampling,
disclosure context, and study methodology.
Child sex. Boys may be less likely than girls to disclose because of a general reluctance to disclose intimate
information, fears related to homosexuality (if the perpetrator is a man; Hunter, 1990; Pescosolido, 1989;
Valente, 2005), or psychological, social, and cultural factors that would lead to overlooking sexual contact as
abuse (if the perpetrator is a woman; Peluso & Putnam,
1996). Some studies have found that boys are less willing to disclose sexual abuse than girls at a forensic interview (e.g., DeVoe & Faller, 1999; Gries, Goh, &
Cavanaugh, 1996; Hershkowitz, Horowitz et al., 2005,
2007), but others find no difference (see DiPietro,
Runyan, & Fredrickson, 1997; Goodman-Brown et al.,
2003; Keary & Fitzpatrick, 1994). Finkelhor et al.’s
(1990) national survey of adults showed that a higher
percentage of adult men, versus women, reported sexual
abuse experiences that they had never disclosed.
Child age at forensic interview. A child’s age affects
his or her ability to understand the abuse, to keep secrets,
and to describe events verbally, all of which can influence the likelihood of disclosure at forensic interviews
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
102 Child Maltreatment
(Brilleslijper-Kater et al., 2004; Pipe et al., 2007).
Research suggests that older, versus younger, children
are more likely to disclose during a forensic interview
(Cantlon, Payne, & Erbaugh, 1996: DiPietro et al., 1997;
Gries et al., 1996: Hershkowitz, Horowitz et al., 2005,
2007; Keary & Fitzpatrick, 1994; Pipe et al., 2007; Sas
& Cunningham, 1995; see London et al., 2005, 2007;
Paine & Hansen, 2002, for reviews). Bradley and Wood
(1996) found no relationship between age and disclosure
at a forensic interview, whereas Wood, Orsak, Murphy,
and Cross (1996) reported that older children made more
credible disclosures. Goodman-Brown et al. (2003),
however, found that older children who disclosed to the
police or social services (which may have occurred outside the context of a forensic interview and at various
locations, including school or home) waited longer to
do so than younger children, with this delay partly
explained by their feeling more responsibility for the
abuse and being more fearful of negative consequences
to others as a result of disclosure. Additionally, the preinvestigation disclosures of preschool age, versus older,
children appear more often to be unplanned and associated with an immediate precipitating event rather than
purposeful (e.g., Campis, Hebden-Curtis, & Demaso,
1993). This may help explain the longer delay until disclosure for older children.
Child age at onset of abuse. Age at onset of abuse may
have an additional relationship to disclosure above and
beyond the age at which the abuse comes to light or
children are interviewed. Child age at onset of abuse may
establish the upper limit of what a child can disclose,
given the child’s developmental abilities at the time of
abuse onset and end (see Brilleslijper-Kater et al., 2004) as
well as given memory limitations associated with the
passing of time (La Rooy, Pipe, & Murray, 2007). Earlier
age of onset might be associated with greater delay
between abuse and forensic interviewing, which can affect
children’s memory retrieval (Salmon & Pipe, 2000).
Many studies make no clear distinction between age
of onset of abuse and age at report of abuse (see, e.g.,
Goodman-Brown et al., 2003; Smith et al., 2000), even
though these two ages can be quite different, can influence disclosure for different reasons, and may have distinct statistical effects on disclosure. We found two
retrospective studies that examined the relationship
between age at onset of sexual abuse and disclosure.
Kogan (2004) analyzed the disclosures of girls ages 12 to
17 who reported unwanted sexual experiences as part of
the National Survey of Adolescents, a national probability study of North American teenagers. In 35% of cases,
girls who were ages 14 to 17 at age of onset had never
disclosed their unwanted sexual experiences compared
to 14% or less for younger ages of onset. Bottoms,
Rudnicki, and Epstein’s (2007) study of a sample of college women found no relationship between age of onset
and disclosure. In several ways, these studies are different from the present one: (a) only small percentages of
examined disclosures were to authority figures (6% and
9%, respectively), (b) neither specifically examined disclosure at a forensic interview, and (c) the sexual abuse
was retrospectively self-reported by the respondent and
by and large uninvestigated. To the best of the authors’
knowledge, no study disentangles and compares the
effect of both age at abuse onset and age at forensic interview on disclosure, despite the different effects they can
be expected to have.
Abuse severity. The severity of abuse could increase
the likelihood of disclosure because the consequences of
the abuse continuing may be perceived as greater and the
need to stop the abuse more critical, and it may be easier
to recognize it as abuse (Cederborg, Lamb, & Laurell,
2007; Lamb & Edgar-Smith, 1994). In addition, children
may less easily recognize as abuse, and remember abuse,
that is relatively minor (e.g., a single episode of fondling;
Cederborg et al., 2007). More severe abuse, however,
might decrease the likelihood of disclosure because the
child might be more afraid of the perpetrator or feel a
heightened sense of self-blame (London et al., 2005). In
their review of the literature, Paine and Hansen (2002)
found that children who had been abused more severely,
as indicated by penetration and physical aggression,
were less likely to disclose. London et al. (2005), however, cite five retrospective studies showing either the
opposite or no relationship between several indices of
severity and disclosure. Findings on the relationship of
abuse duration and frequency, which may also be related
to severity, to disclosure are similarly mixed (see Paine
& Hansen, 2002).
Perpetrator relationship to child. It seems likely that
children abused by extrafamilial perpetrators would
show an increased propensity to disclose their abuse, as
their caregivers might be more supportive of their disclosures and children might feel less loyalty and protectiveness toward the perpetrator. A recent exploratory
study, however, highlights the need to make finer distinctions than “intrafamilial” and “extrafamilial” to
describe the perpetrator’s relationship to the child. In
30 cases of extrafamilial abuse, Hershkowitz, Lanes,
and Lamb (2007) found that most caregivers (75%)
were supportive of their children’s pre-investigation
disclosures only when the perpetrators were strangers;
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 103
when the perpetrator was known to the family, few caregivers were supportive (11%).
Perhaps because of the variability within the categories of “intrafamilial” and “extrafamilial,” research has
been mixed regarding whether the child’s relationship to
the suspect predicts disclosure (see London et al., 2005,
2007). One study examining disclosure at a forensic
interview, however, found that children with an intrafamilial perpetrator disclosed later than children with an
extrafamilial perpetrator (Goodman-Brown et al., 2003).
In this study, it was specifically children’s fear of negative consequences to others that helped explain the
delayed disclosure of intrafamilial abuse.
Suspect age. We know of no research that examines
the probability of disclosure by the age group of the
alleged perpetrator: child, adolescent, or adult. There are
a number of reasons why offender age status could affect
the likelihood of a disclosure. Children may fear, or want
to protect, perpetrators of these different age groups to
different degrees. Additionally, children may feel a different degree of culpability when the abuser is closer in
age to them, may perceive the abuse differently, or may
face different reactions to the abuse when it is first discovered (e.g., that it was natural, mutual sexual exploration) (Kellogg & Huston, 1995).
Caregiver support. Children may worry about the
potential reactions of their caregivers to disclosure (see
Hershkowitz, Lanes et al., 2007), which can range from
being very supportive to being very angry at the child
(Bolen & Lamb, 2004; Deblinger, Steer, & Lippmann,
1999; Elliott & Carnes, 2001; Jensen, Gulbrandsen,
Mossige, Reichelt, & Tjersland, 2005). Of the seven studies we found that examined children’s disclosure of abuse
during a forensic interview (DeVoe & Faller, 1999;
DiPietro et al., 1997; Elliott & Briere, 1994; Gries et al.,
1996; Hershkowitz et al., 2005; Keary & Fitzpatrick, 1994;
Lawson & Chaffin, 1992), only two investigated its relationship to caregiver support. Lawson and Chaffin (1992)
found that children with sexually transmitted infections
whose caregivers were supportive, versus unsupportive,
disclosed at a 3.5 times higher rate, whereas Elliott and
Briere (1994) found that children showed a higher rate of
recanting disclosures of abuse when their caregivers were
unsupportive. In an exploratory study of disclosure of
extrafamilial abuse by children ages 7 to 12, Hershkowitz,
Lanes et al. (2007) found that children’s willingness to disclose immediately and of their own accord decreased when
they expected negative caregiver reactions.
Initial disclosure. An initial disclosure might be
expected to predict a child’s disclosure at a forensic
interview, as it may reflect the child’s readiness to disclose the abuse or decrease children’s fear of disclosing
to an authority figure (see London et al., 2007). Both
Keary and Fitzpatrick (1994) and DiPietro et al. (1997)
studied disclosures among children referred to a
children’s hospital for a sexual abuse evaluation by a
multidisciplinary team and found that a pre-investigation disclosure was highly predictive of a disclosure
during a forensic interview. In Keary and Fitzpatrick’s
(1994) study, 86% of children (ages 3 to 17) who gave
a prior disclosure of abuse disclosed during the forensic interview compared to 14% of children without a
prior disclosure. Similarly, DiPietro et al. (1997) found
that 72% of their sample (range = 1.4 to 22 years of
age; M = 7.5) who gave a prior disclosure disclosed at
the forensic interview versus 7% of children without a
prior disclosure. One study focusing on a different
interview setting and another study focusing on a different population also showed high rates of repeat
disclosure. DeVoe and Faller (1999) found a 93% rate
of repeat disclosure, compared to a 25% rate of firsttime disclosure, from pre-evaluation to interview at a
multidisciplinary clinic among children ages 5 to 10,
although 15% of the repeat disclosures came during
a second or later interview. A sample of foster
children also showed a 93% rate of repeat disclosure
(Gries et al., 1996).
CACs and Child Disclosure
CACs are designed to facilitate child abuse investigation and prosecution while reducing child and family
stress and responding to children’s medical and therapeutic needs (see Cross, Jones, Walsh, Simone, &
Kolko, 2007; Cross et al., in press; Jackson, 2004;
Simone, Cross, Jones, & Walsh, 2005; Walsh, Jones, &
Cross, 2003). CACs might facilitate disclosures if
children find the environment less stressful and more
comforting than other possible interview locations
(e.g., schools, homes, or police stations). The specially
trained child forensic interviewers at CACs may also be
more skilled at obtaining disclosures from children. It
is also possible that CACs, by virtue of their community role, may receive cases after a disclosure has
already been made; therefore, a comparison of disclosure across CAC and non-CAC cases should examine
whether the interviewed children differed with respect
to initial disclosures.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
104 Child Maltreatment
The Present Study
As previous studies have, the present research
explores child, abuse, and suspect characteristics that
predict disclosure at a forensic interview. Expanding on
previous studies, it explores more thoroughly how previous disclosure relates to disclosure at a forensic interview, examines how both age at onset and age at
forensic interview relate to disclosure, and tests whether
disclosure rates differ across communities, including a
comparison of communities with and without a CAC.
Method
This study uses secondary data analysis from a larger
study, the Multi-Site Evaluation of Children’s
Advocacy Centers. This project was designed to evaluate the impact of CACs on child abuse victims, their
families, investigation procedures, and communities.
For additional details on study methodology, see Cross,
Jones, Walsh, Simone, and Kolko (2007) and Cross,
Jones, Walsh, Simone, Kolko, et al. (in press). Research
teams collected data at four sites across the United
States: the Dallas CAC (Dallas, Texas), the Dee Norton
Lowcountry Children’s Center, Inc. (Charleston, South
Carolina), the National CAC (Huntsville, Alabama),
and the Pittsburgh Child Advocacy Center (Pittsburgh,
Pennsylvania).
Sample
Information was collected on a sample of each site’s
CAC cases and on a sample of child abuse cases from
within-state comparison communities without CACs.
Cases were enrolled into the study between December
2001 and December 2003 by research teams at each of
the four sites. In general, the sample included every
available case seen at the CAC and every available case
investigated by the comparison community CPS agencies during the enrollment period. The South Carolina
and Dallas County comparison samples also included
police cases. When there were too many cases to
abstract, a systematic sampling process (e.g., enrolling
each third case) was used. If there were multiple victims
or perpetrators per case, site research staff randomly
selected a “target” subject.
The present study included only cases of alleged sexual
abuse (N = 1,221). Furthermore, to be eligible for analyses on children’s full disclosures during their forensic
interviews, cases had to have included a child forensic
interview with a known outcome (N = 1,101). The sample
was further restricted to cases for which at least one
investigating/evaluating party—law enforcement, CPS,
CAC, or medical—believed that sexual abuse had or may
have occurred, given that an analysis of predictors of
abuse disclosure assumes that abuse occurred. The final
sample consisted of 987 cases, 81% of the entire sample
of sexual abuse cases.
Data Collection
At all sites, case abstractors on the research team collected data from case records. In addition to CAC
records, data abstraction for CAC cases made use of
records from CPS (59% of cases), police (58%), prosecutors (34.5%), and other agencies (e.g., mental health,
medical, and school; 29.5%). Comparison cases used
records from CPS (68% of cases), police (66.5%), prosecutors (26.5%), and other agencies (30%).
Variables
Case abstractors coded information on each forensic
interview. We defined forensic interview as a professional
interview designed to assess or evaluate the truth about a
suspicion of child maltreatment (as well as identify the
who, what, where, and when of the abuse). We excluded
the following situations: any pre-investigation disclosures
of abuse by the child to a parent, teacher, friend, and so
on; talks that a parent had with a child to better understand what happened; a discussion of the abuse by a mental health professional for clinical purposes; and any
initial contact by CPS or the police to assess briefly
immediate risk (e.g., a minimal facts interview). Forensic
interviewers at the CACs had a bachelor’s or master’s
degree and social work, counseling, or child welfare
backgrounds. They followed American Professional
Society on the Abuse of Children interviewing guidelines.
When research staff lacked the interview content to code
a police or CPS interview, they erred on the side of calling any fact-finding interview with a CPS worker or
police officer a “forensic interview.”
For the disclosure variable, case abstractors recorded
whether the child denied, disclosed fully or partially, or
recanted allegations of abuse for each forensic interview (see Cross, Jones, Walsh, Simone, & Kolko, 2007;
Cross et al., in press). Children’s full disclosures were
defined as a disclosure over the course of any single
forensic interview of all sexual activity (all known sexual acts, all known incidents) that came to be known
throughout the investigation, whether by the child’s
own disclosure, the suspect’s confession, or a witness’s
account.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 105
A range of demographic, family, abuse, and suspect
variables were also coded. In the present analysis, the
following variables were used as predictors of disclosure: whether there was a pre-interview disclosure
(described within CPS, police, or medical records),
whether a previous disclosure was the investigation elicitor, child age at onset of abuse, child age at forensic
interview, child sex, abuse severity (vaginal or anal penetration, as reported by the child, suspect or eyewitness),
alleged perpetrator relationship to child (intrafamilial or
extrafamilial, as the study measures limited our ability to
make finer distinctions; intrafamilial included biological,
adoptive, and step-parents; siblings and stepsiblings; and
other relatives; extrafamilial consisted of parents’ paramours, foster parents, nonrelative caregivers, nonrelative
authority figures, and known and unknown adults and
children), alleged suspect age, caregiver support, and
interview setting (CAC or non-CAC). Nonoffending
caregiver support for the child was coded as yes versus
no/ambivalent according to CPS’ and police investigators’ records of their judgment of the caregiver’s supportiveness (judged according to the degree to which the
caregiver believed the child’s disclosure and was willing
to stay with the child and prevent the suspect’s having
contact with the child). To avoid relying only on investigators’ judgments, we also examined caregiver knowledge of the abuse prior to the investigation and specific
caregiver actions related to support. We lacked the
resources to conduct a formal reliability assessment on
variables analyzed; however, the variables here mostly
represent concrete events, and investigation records contained specific documentation on these events.
Data Analysis
Bivariate analyses using Pearson χ2 and t tests were
conducted to identify variables that distinguished full
disclosers from nonfull disclosers (which included partial disclosures, no disclosures, and denials). The potential predictors of disclosure we examined were chosen
because previous research or practice theory suggested
their contribution. The variables that differed between
the study’s groups (full vs. nonfull disclosers) were then
entered into binary logistic regression equations to determine their unique association with disclosure.
In the logistic regression equation, we replaced missing data with the sample mean of the continuous variables with at least 5% missing data (age of abuse onset
and age at forensic interview) and adding “missing” as a
category for the categorical variables with at least 5%
missing data (pre-interview disclosure, caregiver support, abuse severity, child-suspect relationship) (Cohen,
Cohen, West, & Aiken, 2003). When the value of the
sample mean for age of abuse onset was higher than the
child’s age at first interview or when the value of the
sample mean for age at first interview was higher than
the child’s age during the investigation, no replacement
occurred (consequently, replacement occurred for a total
of 12.6% of the age of abuse onset data and 4% of the
age at forensic interview data). The logistic regression
analysis then included a variable for age of abuse onset
that represented whether the case originally had a missing value; this corrects for any bias that might arise from
mean substitution (see Cohen et al., 2003). To account
for possible curvilinear effects of age (e.g., both very
young children and adolescents having lower rates than
middle-age children), we entered a squared version of
age to test for a quadratic effect. Age variables were centered to reduce multicollinearity (see Cohen et al., 2003).
Sample Characteristics
Table 1 shows sample characteristics. The majority of
children were White (Hispanic and non-Hispanic). Of
the children identified as non-White, the majority were
identified as Black (28%); the remaining were either
biracial (4%), Asian (0.3%), or Pacific Islander (0.2%).
One was identified as Native American and the race of
eight (.8%) children was unknown. The mean age at
onset of abuse was 8.90 (minimum = 1) and the mean
age at forensic interview was 9.91 (minimum = 2). In
55% of cases (n = 921), the suspect was a family
member—defined as a blood relative, a stepparent, or an
adoptive parent. In 30% of cases (n = 941), the suspect
and child were living together when the abuse investigation began. When there was cohabitation, the suspect
was usually related to the child (71%; n = 293).
Child age at abuse onset and at forensic interview
were highly positively correlated, r(680) = 0.88, p < .01.
The average difference between these two ages was 1.47
years, but the distribution of this variable was skewed:
65% of the sample had the forensic interview within 1
year of onset, 29% between 1 year to 7 years after onset,
and 4% more than 7 years after onset. Children had an
average of 1.4 forensic interviews (SD = 0.65, n = 979).
Results
Full disclosures occurred for 73% (722) of cases, partial disclosures for 12% (117), no disclosures (100) for
10%, and denials for 5% (48) of all cases. Of full disclosers with available time data (n = 482), 43% disclosed
the abuse months after the last episode, 28% disclosed
days after the last episode, and 29% within hours (see
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
106 Child Maltreatment
Table 1
Study Sample Case Data (N = 987)
Variable
Child
Female
White
Age at abuse onset
Age at first interview
Abuse
Penetration
Lasted beyond a week
Suspect
Intrafamilial
Cohabiting with child
Male
White
Age
Disclosure
Full
Years from onset
Disclosed within hours
of last abusea
Prior disclosure
Caregiver supportive
Agency/community
CAC
Yes
No
State
Pennsylvania
Texas
South Carolina
Alabama
N
Percentage
of N
Figure 1
Disclosure Rates by Study Communities
M
SD
100%
90%
985
979
842
969
81
67
837
542
33
44
921
941
943
873
977
55
30
93.5
66.5
987
718
482
73
29
906
761
62
83
677
310
69
31
141
317
319
210
14
32
32
21
80%
8.90
9.91
3.74
4.06
70%
60%
50%
40%
30%
20%
10%
32.97
17.69
0%
Alabama
1.64
Pennsylvania
2.14
*p = .03. **p < .001.
a. Versus disclosing within days or months of last abuse.
Table 1). The full disclosure rate varied across communities from 61% to 89% (see Figure 1).
Bivariate Relationship of Predictor Variables
and Child Disclosure
Tables 2 and 3 show bivariate relationships of categorical and continuous predictor variables, respectively, to
child disclosure. Girls were significantly more likely to
make a full disclosure than boys. A full disclosure was
positively related to both child age at abuse onset and
child age at forensic interview. For both age at onset and
age at forensic interview, only a small majority of
children 0 to 6 years of age disclosed. Disclosure rates
were much higher for older groups. The age of onset for
disclosers (M = 9.67) was on average 3 years older than
for nondisclosers (M = 6.22), t(734) = –10.45, p < .001,
and disclosers’ age at forensic interview (M = 10.85) was
CAC
South
Carolina
Texas
Comparison
almost 3 years older on average compared to nondisclosers (M = 7.34), t(967) = –12.90, p < .001. Suspect
age was negatively related to disclosure: suspects of
children giving full disclosures (M = 32.12) were more
than 3 years younger than suspects of non-full-disclosing
children (M = 35.50), t(382.42) = 2.24, p < .03.
Children with an extrafamilial relationship with suspects were slightly more apt to make a full disclosure,
but the child-suspect relationship was significantly correlated with several other predictor variables (child age
at abuse onset, r(730) = 0.20, p < .001, child age at interview, r(904) = 0.13, p < .001, and suspect age, r(661) =
0.26, p < .001. A higher percentage of children disclosing severe abuse (actual/attempted vaginal or anal intercourse) gave full disclosures as compared to those with
less severe abuse.
Sixty percent of cases without a pre-investigation disclosure had a full disclosure at the forensic interview
compared to 81% of cases that had a prior disclosure,
χ2(1, n = 906) = 47.48, p < .001. Nonoffending caregiver
support for the child, as coded from investigators’
records, was only marginally related to disclosure (p =
.08). Nevertheless, several specific caregiver supportive
actions were related to children’s full disclosures during
forensic interviews (see Table 2): contacting law
enforcement, contacting another, restricting suspect contact with the child, and removing the suspect. No significant difference was associated with caregivers’
contacting CPS, a CAC, a medical professional, relocating the child, or relocating with the child.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 107
Table 2
Disclosure Rates by Categorical Predictor Variables
Variables
Child sex
Female
Male
Child race
White
Minority
Child age at onset
0 to 6
7 to 12
13 to 17
Child age at interview
2 to 6
7 to 12
13 to 17
Suspect-child relationship
Intrafamilial
Extrafamilial
Suspect cohabitation with child
Yes
No
Vaginal/anal penetration
Yes
No
Duration of abuse
More than 1 week
Less than 1 week
Previous disclosure
Yes
No
Caregiver support
Yes
No
Caregiver knowledge of abuse
Aware
Unaware
Suspected it, but uncertain
Caregiver response to abuse
Yes, contacted police
Yes, contacted others
Yes, restricted
child-suspect contact
Yes, removed the suspect
Elicitor of abuse investigation
Pre-investigation disclosure
Witness to the abuse
Child’s behavior or symptoms
Physical signs/medical evidence
Other
Children’s Advocacy Center
Yes
No
Location of communities (state)
Pennsylvania
Texas
South Carolina
Alabama
a. p = .08.
*p < .04. **p < .01. ***p < .001.
N
% Disclosed
χ2
Table 3
Age Variables and Years From Abuse Onset for
Disclosure and Nondisclosure Cases
Variable
985
76
62
15.30***
979
73
74
0.18
842
58
83
92
79.85***
969
51.5
75
92
117.19***
921
70
78
7.77**
70
75
2.02
669
837
89
77
18.44***
542
89
85
1.56
906
81
60
47.48***
761
76
69
3.10a
83
78
62
8.00*
83.5 (vs. 72 no)
86 (vs. 75 no)
82 (vs. 75 no)
13.92***
8.65**
4.32*
91 (vs. 77 no)
4.37*
801
607
906
81
71
59
54
57
52.73***
987
71
78
5.28a
62
73
75
78
11.24**
987
Age at abuse onset
Age at first interview
Suspect age
Years from onset
Disclosure Nondisclosure
9.67
10.85
32.12
1.59
6.22
7.34
35.30
1.80
t
–10.45**
–12.90**
2.24*
0.31
df
734
967
382.42
716
*p < .05. **p < .001.
The rate of disclosure was significantly different
between CACs (71% full) and comparison sites (78%
full), χ2(1, N = 987) = 4.85, p = .03. We also examined
differences with respect to disclosure when disclosure
was divided into three categories and found a marginally
significant difference between CACs (71% full, 13%
partial, and 16% no disclosure or denial) and comparison
sites (78% full, 9% partial, and 13% no disclosure or
denial), χ2(2, N = 987) = 5.28, p = .07. Though there was
a significant difference between CAC and comparison
communities, they differed significantly on child race,
age at abuse onset, age at forensic interview, and the perpetrator relationship to the child; they showed no difference with respect to the percentage of children giving
pre-interview disclosures (62% for both), χ2(1, n = 906)
= 0.11, p > .10. There were significant differences by
state, although this almost certainly reflects specific
characteristics of a particular site versus characteristics
of the state as a whole. In the Pennsylvania communities,
for example, where lower disclosure rates were found, a
number of the children were interviewed at the Emergency
Department of a Children’s Hospital.
We also examined correlations between predictor
variables and disclosure and among study variables
(see Table 4). Correlations between child race, suspect
race, and other variables were also examined separately. Child race (minority) was positively correlated
with having an interview at a CAC, r(979) = .20, p <
.001, suspect race, r(870) = .80, p < .001, and
vaginal/anal penetration, r(829) = .12, p = .001. It was
negatively correlated with caregiver support, r(753) =
–.09, p = .02, and a disclosure being the investigation
elicitor, r(977) = –.07, p < .02. Suspect race (minority)
was similarly related to the foregoing variables, except
disclosure being the investigation elicitor, for which it
had no significant relationship (p > .10). Suspect race
(minority) was uniquely correlated with suspect age,
r(872) = –.08, p = .004, and extrafamilial abuse, r(873) =
.10, p = .004.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
108
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
1.0
.13**
985
.36**
736
.38**
969
.07*
980
0.06
669
.15**
837
.29**
980
.19**
985
0.06
761
.14**
740
.11**
714
.08*
674
.09*
607
.07*
985
0.05
977
0.02
824
—
1.0
985
.18**
734
.16**
967
0.03
979
0.06
667
.07*
835
0.03
978
0.06
983
.10**
760
0.01
738
0.01
712
0.02
672
0.03
605
0.04
985
.13**
976
0.04
822
2
—
1.0
736
.86**
718
.20**
730
0.04
565
.22**
684
.24**
729
.12**
734
0.05
609
0.01
607
.09*
589
0.06
550
0.02
494
.12**
736
.11**
727
.29**
718
1.0
969
.11**
963
0.02
657
.29**
822
.24**
964
.15**
967
.13**
743
–.01
723
.10*
697
.09*
657
0.01
590
.16**
969
0.02
960
.34**
824
—
—
4
—
—
3
1.0
980
.25**
666
0.05
833
0.06
975
.16**
978
0.05
758
.11**
737
.10**
711
.10**
671
0.1
604
.12**
980
.08*
977
.09*
819
—
—
—
—
5
—
1.0
669
.11*
599
.21**
666
0.02
667
.24**
555
.14**
534
0.04
505
0.07
466
.20**
401
0.05
669
.10*
663
.12**
593
1.0
837
.11**
837
.08*
747
.10*
684
.09*
699
–.03
681
0.05
642
0.01
586
0.01
837
.18**
830
.08*
747
—
—
—
—
—
7
—
—
—
—
—
6
1.0
980
0.02
978
0.07
756
.22**
738
0.01
712
0.02
672
0.01
605
0.01
980
.15**
972
.19**
819
—
—
—
—
—
—
—
8
1.0
985
.11**
761
0.03
739
0.01
713
0.04
673
0.01
606
.12**
985
.09**
816
0.01
822
—
—
—
—
—
—
—
—
9
1.0
761
.22**
648
0.03
632
.15**
588
0.01
530
0.02
761
.16**
757
.13**
658
—
—
—
—
—
—
—
—
—
10
—
—
—
—
—
—
—
—
—
—
12
1.0
—
740
.08*
1.0
701
714
.17**
.32**
662
657
0.04
.21**
601
600
0.03
0.002
740
714
0.02
0.02
735
709
0.01
0.04
654
632
—
—
—
—
—
—
—
—
—
—
11
Note: Spearman’s rho was used except when both variables were continuous; then, Pearson’s r was used. Bold signifies a negative relationship.
*p ≤ .05. ** p ≤ .01.
17. Years from onset
16. Suspect age
15. CAC
14. Caregiver remove suspect (yes)
13. Caregiver restrict contact (yes)
12. Caregiver contact others (yes)
11. Caregiver contact police (yes)
10. Caregiver support (yes)
9. Disclosure investigation elicitor
8. Abuse duration (> 1 week)
7. Vaginal/anal penetration (yes)
5. Suspect-child relationship
(extrafamilial)
6. Suspect cohabit (no)
4. Child age at forensic interview
1. Full disclosure
2. Child sex
(female)
3. Child age at abuse onset
1
Table 4
Intercorrelations Between Study Variables
—
—
—
—
—
—
—
—
—
—
—
—
14
1.0
—
674
.25**
1.0
594
607
0.02
0.06
674
607
0.03
0.003
669
602
.14** 0.02
592
534
—
—
—
—
—
—
—
—
—
—
—
—
13
1.0
985
–.06*
977
0.03
824
—
—
—
—
—
—
—
—
—
—
—
—
—
—
15
1.0
977
.10**
816
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
16
Lippert et al. / Telling Interviewers About Sexual Abuse 109
Table 5
Final Logistic Regression Predicting Child Full Disclosure (N = 739)
Predictor
Child sex (female vs. male)
Age of abuse onset
Age variable
Missing age of onset (yes/no)
Age at forensic interview
Suspect age
Suspect relationship to child (intra v. extra)
Vaginal/anal penetration
Yes vs. no
Missing severity (yes/no)
Previous disclosure
Yes vs. no
Missing previous disclosure (yes/no)
Caregiver support
Yes vs. no
Missing caregiver support (yes/no)
State
Pittsburgh vs. Huntsville
Dallas vs. Huntsville
Charleston vs. Huntsville
Β
SE
Wald
p
Odds Ratio
95% CI
0.55
.26
4.37
.037
1.73
1.04 to 2.89
0.19
–1.08
0.10
–0.01
–0.01
.06
.43
.05
.01
.01
10.95
6.43
3.75
0.60
1.35
.001
.011
.053
.437
.245
1.21
0.34
1.11
1.00
0.99
1.08 to 1.35
0.15 to 0.78
1.00 to 1.23
0.99 to 1.01
0.98 to 1.01
0.24
–2.61
.26
.53
0.79
24.07
.373
< .001
1.26
0.07
0.75 to 2.12
0.03 to 0.21
0.83
1.60
.23
.81
13.36
3.87
< .001
.049
2.30
4.96
1.47 to 3.59
1.01 to 24.44
1.15
0.01
.30
.32
14.70
0.01
< .001
.970
3.14
1.01
1.75 to 5.65
0.54 to 1.88
0.13
0.87
2.05
.31
.28
.44
0.18
9.40
22.10
.673
.002
< .001
1.14
2.38
7.74
0.62 to 2.10
1.37 to 4.15
3.30 to 18.17
Note: CI = Confidence Interval.
*p < .05. **p < .005.
Multiple Predictor Model
In general, variables demonstrated by binary analyses
to be significantly related to disclosure were then entered
simultaneously into a logistic regression model. There
were two exceptions. First, the variable “elicitor of abuse
investigation” (disclosure vs. nondisclosure) was left out
of the regression model because it is closely related to
whether a child had already disclosed prior to the forensic interview. Second, caregiver support was entered into
the regression. Although caregiver support was only
marginally significant at the bivariate level of analysis,
several variables measuring actions related to caregiver
support were significant bivariate predictors of disclosure. These other caregiver actions, however, were each
too specific and had too many missing data points to be
good candidates for the logistic regression model.
Table 5 shows the model, which correctly classified
84% of the cases, 96% of the full disclosers, and 39% of
the nonfull disclosers. Child sex, child age at abuse
onset, abuse severity, pre-investigation disclosure, caregiver support at the time of the pre-interview disclosure,
and investigation location each were uniquely associated
with children’s full disclosures. The adjusted odds of full
disclosure during a forensic interview were 1.7 times
greater for girls than boys. The adjusted odds of full disclosure were also 1.3 times greater per increased year of
age at the time of abuse onset. Children whose severity
of abuse was unknown had 0.7 the adjusted odds of full
disclosure as children disclosing nonsevere abuse. This
severity effect, however, likely reflects the reality that
when children give partial or no disclosures of abuse,
data on severity are often missing.
When there was a pre-investigation disclosure, the
adjusted odds of disclosure at forensic interview were 2.3
times the adjusted odds of disclosure when there was no
pre-investigation disclosure. Children whose primary
nonoffending caregivers were judged by at least one investigator to be supportive had 3.14 the adjusted odds of disclosure at the forensic interview as children without a
supportive caregiver. Finally, the Texas communities had
adjusted odds of disclosure that were 2.38 greater than the
Alabama communities’, and the South Carolina communities had adjusted odds of disclosure that were 7.74 times
greater than the Alabama communities’.
Because, as Figure 1 suggests, disclosure rates varied
across individual communities within the CAC and
state groupings, an additional logistic regression model
was run that added a State × CAC interaction effect. In
the additional model, this interaction effect was statistically significant (p = .03), but the model with the
interaction effect was inferior to the prior model
because of multicollinearity between “state” and the
interaction variable.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
110 Child Maltreatment
Discussion
Unlike previous studies, the present research illuminates the specific relationship of child age at onset and
child age at forensic interview to disclosure and suggests
that rates of disclosure at the forensic interview may vary
across communities. These results were consistent across
bivariate analyses and a logistic regression that statistically controlled for relationships among predictor variables. The present study also replicates previous studies
on the relationship of child sex and age, previous disclosure, and caregiver support to disclosure. Consistent with
previous research (e.g., DeVoe & Faller, 1999; Gries et
al., 1996; Hershkowitz et al., 2005), girls had a higher
rate of full disclosure during a forensic interview than
boys. The finding that children who disclosed prior to the
forensic interview more often disclosed at the forensic
interview than did children who had made no disclosure
prior to the interview also replicated results of previous
research (DiPietro et al., 1997; Keary & Fitzpatrick,
1994). The positive relationship between caregiver support and child disclosure is consistent with studies
reviewed by Paine and Hansen (2002). Nonetheless, the
fact that the investigators documenting caregiver support
knew whether children disclosed should make us cautious about interpreting results for this variable. More
research on caregiver support that operationalizes support through caregivers’ actions is needed because of the
biases introduced by relying on the judgments of investigators. The present study outlined various actions that
could be used to define caregiver support. In bivariate
relationships, contacting law enforcement, contacting
another, restricting child-suspect contact, and removing
the suspect were related to an increased likelihood of a
full disclosure, but because of substantial missing data,
these variables were excluded from the regression analysis.
These variables merit further examination. In bivariate
relationships, disclosure was more common among cases
involving extrafamilial abuse, vaginal or anal penetration,
and older suspects. But these variables were no longer
significant when we controlled for related variables such
as location and child age at onset and interview.
Age of Onset and Age at Forensic Interview
The findings on child age replicate and extend findings of previous studies by repeating the relationship of
age at interview to disclosure but showing that age at
onset has a unique association to disclosure at a forensic
interview as well. As previous studies (DiPietro et al.,
1997; Hershkowitz et al., 2005; Keary & Fitzpatrick,
1994; see London et al., 2005, 2007; Paine & Hansen,
2002, for reviews) have found, children who were older
at the time of the forensic interview were likelier to disclose. Age at interview, however, was correlated with age
of onset, and when a multiple predictor model including
both ages was examined, both age of onset and age at
interview were independently related to disclosure, with
age of onset having a stronger independent relationship.
The relationship of age of onset to disclosure may be
explained by the reality that the earlier that the abuse
began, the older at least some of the memories of abuse
will be. Children’s ability to disclose may be hampered
by the limits of their memory and developmental constraints on their capacity to understand what happened to
them at the time. Moreover, a number of other variables
(e.g., family dysfunction) may be correlated with age of
onset of abuse, which may also affect disclosure. In a
bivariate analysis, age at forensic interview was significantly related to disclosure but was a marginally statistically significant predictor (p = .05) when its effect
independent of age at onset was estimated by the logistic
regression analysis. It is likely that older children’s
greater ability to understand what is needed at a forensic
interview and generally to relay more comprehensive
narratives (e.g., Orbach & Lamb, 2007) also increases
the probability of a full disclosure, independently of
when the abuse happened. The logistic regression equation reduced the statistical power of both age at onset and
age at forensic interview as predictors given the substantial multicollinearity between them. Nonetheless, it
allowed estimation of the unique effect of each age on
disclosure to help illuminate what age-related processes
might affect disclosure.
Differences by Community
A lesson from the present study is the dependence of
disclosure rates on the particular child population and
organizations studied. The present study’s disclosure
rates from abuse investigations by CACs, CPS, and
police were substantially higher than disclosure rates of
studies that exclusively examined disclosure among
children referred to a children’s hospital (DiPietro et al.,
1997; Keary & Fitzpatrick, 1994). In the present study,
too, disclosure rates differed significantly between communities. This seems to be a function of particular characteristics of the multiple communities rather than
attributable to differences by state or by CAC status. The
CAC status variable was nonsignificantly related to disclosure, and the main effect of state was probably a function of the State × CAC interaction effect. Despite reason
to believe that CACs might facilitate child disclosure, the
rates of full child disclosures were similar between
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 111
CACs and non-CACs, once we accounted for case differences. Also, it should be noted that children interviewed
at CACs and non-CACs showed an identical rate of preinterview disclosure. Further research should examine
the influence of setting characteristics, interviewers’
background (forensic interviewer vs. CPS vs. police),
interview methods, the diversity of cases, and methodological differences on disclosure rates. For the time
being, there is a limited ability to estimate the probability that, given a particular setting, a given child victim
will disclose sexual abuse.
Conclusion
District attorneys often remind juries that there are
usually only two witnesses to the crime of sexual abuse:
the victim and the suspect. A disclosure during a forensic interview is therefore often critical to an effective
response to child sexual abuse. For children whose investigation begins for reasons other than a disclosure, our
finding that most of them disclosed fully during a forensic interview is encouraging. At the same time, there are
lessons here that might help child abuse investigators,
evaluators, and researchers understand better the conditions that help children disclose. Investigators need to be
aware of whether children disclosed previously and what
circumstances and responses surrounded their disclosures. They need to be sensitive to how old children were
when abuse began and how long children have held on to
the secret. Knowing the relevance of parental support to
child disclosure, they need to assist parents to support
their children. Research is needed to understand how a
particular setting, its methods, and its child population
might affect disclosure. Perhaps most of all, child abuse
investigators and evaluators should have confidence that
they can assist most child victims to disclose sexual
abuse under the right conditions.
References
Bolen, R., & Lamb, J. L. (2004). Ambivalence of nonoffending
guardians after child sexual abuse disclosure. Journal of
Interpersonal Violence, 19, 185-211.
Bottoms, B. L., Rudnicki, A. G., & Epstein, M. A. (2007). A retrospective study of factors affecting the disclosure of childhood sexual and physical abuse. In M. Pipe, M. E. Lamb, Y. Orbach, & A.
Cederborg (Eds.), Child sexual abuse: Disclosure, delay and
denial (pp. 175-194). Mahwah, NJ: Lawrence Erlbaum.
Bradley, A. R., & Wood, J. M. (1996). How do children tell? The disclosure process in child sexual abuse. Child Abuse & Neglect, 9,
881-891.
Brilleslijper-Kater, S. N., Friedrich, W. N., & Corwin, D. L. (2004).
Sexual knowledge and emotional reaction as indicators of sexual
abuse in young children: Theory and research challenges. Child
Abuse & Neglect, 28, 1007-1017.
Campis, L. B., Hebden-Curtis, J., & Demaso, D. R. (1993).
Developmental differences in detection and disclosure of sexual
abuse. Journal of the American Academy of Child and Adolescent
Psychiatry, 32, 920-924.
Cantlon, J., Payne, G., & Erbaugh, C. (1996). Outcome-based practice: Disclosure rates of child sexual abuse comparing allegation
blind and allegation informed structured interviews. Child Abuse
& Neglect, 20, 1113-1120.
Cederborg, A., Lamb, M. E., & Laurell, O. (2007). Delay of
disclosure, minimization, and denial of abuse when the evidence
is unambiguous: A multivictim case. In M. Pipe, M. E. Lamb, Y.
Orbach, & A. Cederborg (Eds.), Child sexual abuse: Disclosure,
delay and denial (pp. 159-173). Mahwah, NJ: Lawrence Erlbaum.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences
(3rd ed.). Mahwah, NJ: Lawrence Erlbaum.
Cross, T. P., Jones, L. J., Walsh, W., Simone, M., & Kolko, D. J. (2007).
Child forensic interviewing in Children’s Advocacy Centers: Empirical
data on a practice model. Child Abuse & Neglect, 31, 1031-1052.
Cross, T. P., Jones, L. J., Walsh, W., Simone, M., Kolko, D. J.,
Szczepanski, J., et al. (in press). The Multi-Site Evaluation of
Children’s Advocacy Centers: Overview of the results and implications for practice (OJJDP Crimes Against Children Series)
(research bulletin). Washington, DC: Office of Juvenile Justice
and Delinquency Prevention.
Deblinger, E., Steer, R., & Lippmann, J. (1999). Maternal factors
associated with sexually abused children’s psychosocial adjustment. Child Maltreatment, 4, 13-19.
DeVoe, E. R., & Faller, K. C. (1999). The characteristics of disclosure
among children who may have been sexually abused. Child
Maltreatment, 4, 217-227.
DiPietro, E. K., Runyan, D. K., & Fredrickson, D. D. (1997).
Predictors of disclosure during medical evaluation for suspected
sexual abuse. Journal of Child Sexual Abuse, 6, 133-142.
Elliott, A. N., & Carnes, C. N. (2001). Reactions of nonoffending
parents to the sexual abuse of their child: A review of the literature. Child Maltreatment, 6, 314-331.
Elliott, D. M., & Briere, J. (1994). Forensic sexual abuse evaluations
of older children: Disclosures and symptomatology. Behavioral
Sciences and the Law, 12, 261-277.
Finkelhor, D., Cross, T. P., & Cantor, E. (2005). The justice system
for juvenile victims: A comprehensive model of case flow.
Trauma, Violence, & Abuse, 6, 83-102.
Finkelhor, D., Cross, T., & Cantor, E. (2006). How the justice system
responds to juvenile victims: A comprehensive model (Juvenile
Justice Bulletin—NCJ210951). Washington, DC: Office of
Juvenile Justice and Delinquency Prevention.
Finkelhor, D., Hotaling, G., Lewis, I., & Smith, C. (1990). Sexual
abuse in a national survey of adult men and women: Prevalence,
characteristics, and risk factors. Child Abuse & Neglect, 14, 14-28.
Goodman-Brown, T. B., Edelstein, R. S., Goodman, G. S., Jones, D. P. H.,
& Gordon, D. S. (2003). Why children tell: A model of children’s
disclosure of sexual abuse. Child Abuse & Neglect, 27, 525-540.
Gries, L. T., Goh, D. S., & Cavanaugh, J. (1996). Factors associated
with disclosure during child abuse assessment. Journal of Child
Sexual Abuse, 5(3), 1-19.
Hershkowitz, I., Horowitz, D., & Lamb, M. (2005). Trends in children’s
disclosure of abuse in Israel: A national study. Child Abuse &
Neglect, 29, 1203-1214.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
112 Child Maltreatment
Hershkowitz, I., Horowitz, D., & Lamb, M. (2007). Individual and
family variables associated with disclosure and nondisclosure of
child abuse in Israel. In M. Pipe, M. E. Lamb, Y. Orbach, & A.
Cederborg (Eds.), Child sexual abuse: Disclosure, delay and
denial (pp. 63-75). Mahwah, NJ: Lawrence Erlbaum.
Hershkowitz, I., Lanes, O., & Lamb, M. E. (2007). Exploring the disclosure of child sexual abuse with alleged victims and their
parents. Child Abuse & Neglect, 31, 111-123.
Hunter, M. (1990). Abused boys: The neglected victims of sexual
abuse. Lexington, MA: Lexington Books.
Jackson, S. (2004). A USA survey of program services provided by
child advocacy centers. Child Abuse & Neglect, 28, 411-421.
Jensen, T. K., Gulbrandsen, W., Mossige, S., Reichelt, S., &
Tjersland, O. A. (2005). Reporting possible sexual abuse: A qualitative study on children’s perspectives and the context for disclosure. Child Abuse & Neglect, 29, 1395-1413.
Keary, K., & Fitzpatrick, C. (1994). Children’s disclosure of sexual
abuse during formal investigation. Child Abuse & Neglect, 18(7),
543-548.
Kellogg, N. D., & Huston, R. L. (1995). Unwanted sexual experiences in adolescents: Patterns of disclosure. Clinical Pediatrics,
34, 306-312.
Kogan, S. M. (2004). Disclosing unwanted sexual experiences:
Results from a national sample of women. Child Abuse & Neglect,
28, 147-165.
Lamb, S., & Edgar-Smith, S. (1994). Aspects of disclosure mediators
of outcome of childhood sexual abuse. Journal of Interpersonal
Violence, 9, 307-326.
La Rooy, D., Pipe, M.- E., & Murray, J. E. (2007). Enhancing
children’s event recall after long delays. Applied Cognitive
Psychology, 21, 1-17.
Lawson, L., & Chaffin, M. (1992). False negatives in sexual abuse
interviews. Journal of Interpersonal Violence, 7, 532-542.
London, K., Bruck, M., Ceci, S. J., & Shuman, D. W. (2005).
Disclosure of child sexual abuse: What does the research tell us
about the ways that children tell? Psychology, Public Policy, &
Law, 11, 194-226.
London, K., Bruck, M., Ceci, S. J., & Shuman, D. W. (2007).
Disclosure of child sexual abuse: A review of the contemporary
empirical literature. In M. Pipe, M. E. Lamb, Y. Orbach, & A.
Cederborg (Eds.), Child sexual abuse: Disclosure, delay and
denial (pp. 11-39). Mahwah, NJ: Lawrence Erlbaum.
Orbach, Y., & Lamb, M. E. (2007). Young children’s references to
temporal attributes of allegedly experienced events in the course
of forensic interviews. Child Development, 78(4), 1100-1120.
Paine, M. L., & Hansen, D. J. (2002). Factors influencing children to
self-disclose sexual abuse. Clinical Psychology Review, 22, 271-295.
Peluso, E., & Putnam, M. (1996). Case study: Sexual abuse of boys
by females. Journal of the American Academy of Child & Adolescent
Psychiatry, 35, 51-54.
Pence, D., & Wilson, C. (1994). Team investigation of child sexual
abuse. Thousand Oaks, CA: Sage.
Pescosolido, F. J. (1989). Sexual abuse of boys by males: Theoretical
and treatment implications. In S. Sgroi (Ed.), Vulnerable
populations: Sexual abuse treatment for children, adult survivors,
offenders, and persons with mental retardation (Vol. 2., pp. 85109). Lexington, MA: Lexington Books.
Pipe, M., Lamb, M. E., Orbach, Y., Stewart, H., Sternberg, K. J., &
Esplin, P. (2007). Factors associated with nondisclosure of suspected abuse during forensic interviews. In M. Pipe, M. E.
Lamb, Y. Orbach, & A. Cederborg (Eds.), Child sexual abuse:
Disclosure, delay and denial (pp. 77-96). Mahwah, NJ: Lawrence
Erlbaum.
Salmon, K., & Pipe, M. (2000). Recalling an event one year later: The
impact of props, drawing and a prior interview. Applied Cognitive
Psychology, 14, 99-120.
Sas, L. D., & Cunningham, A. H. (1995). Tipping the balance to tell
the secret: The public discovery of child sexual abuse. London,
Ontario, Canada: London Family Court Clinic.
Saywitz, K. J., Esplin, P., & Romanoff, S. (2007). A holistic approach
to interviewing and treating children in the legal system. In M.
Pipe, M. E. Lamb, Y. Orbach, & A. Cederborg (Eds.), Child sexual abuse: Disclosure, delay and denial (pp. 219-250). Mahwah,
NJ: Lawrence Erlbaum
Sgroi, S., Blick, L., & Porter, F. (1982). A conceptual framework for
child sexual abuse. In S. Sgroi (Ed.), Handbook of clinical intervention in child sexual abuse (pp. 9-37). Lexington, MA:
Lexington Books.
Simone, M., Cross, T., Jones, L., & Walsh, W. (2005). Children’s
advocacy centers: Understanding the impact of a phenomenon. In
K. Kendall-Tackett & S. Giacomoni (Eds.), Victimization of
children and youth: Patterns of abuse, response strategies (pp.
22-1–22-24). Kingston, NJ: Civic Research Institute, Inc.
Smith, D. W., Letourneau, E. J., Saunders, B. E., Kilpatrick, D. G.,
Resnick, H. S., & Best, C. L. (2000). Delay in disclosure of childhood rape: Results from a national survey. Child Abuse & Neglect,
24(2), 273-287.
Sorenson, T., & Snow, B. (1991). How children tell: The process of
disclosure in child sexual abuse. Child Welfare, 70, 3-15.
Valente, S. (2005). Sexual abuse of boys. Child and Adolescent
Psychiatric Nursing, 18, 10-16.
Walsh, W., Jones, L., & Cross, T. P. (2003). Children’s advocacy centers: One philosophy, many models. APSAC Advisor, 15(3), 3-7.
Wood, B., Orsak, J., Murphy, M., & Cross, H. J. (1996). Semistructured
child sexual abuse interviews: Interview and child characteristics related to credibility of disclosure. Child Abuse & Neglect,
20, 81-92.
Tonya Lippert, PhD, MSSW, coordinates a project involving foster youth at risk of homelessness at New Avenues for
Youth and works for Kaiser Permanente conducting medical
evaluations of child abuse at CARES NW. She worked as a
research manager for the Dallas Children’s Advocacy Center
and the Multi-Site Evaluation of Children’s Advocacy
Centers. Her research interests include understanding how
children disclose child maltreatment, predicting confessions
among child abusers, and identifying effective methods of
engaging and helping at-risk youth and families.
Theodore P. Cross, PhD, is a clinical psychologist and now
visiting research professor at the Children and Family
Research Center in the School of Social Work at the
University of Illinois at Urbana-Champaign. He currently
conducts research on placement outcomes, service delivery
and well-being for children in foster care. Over the last 23
years, his research has focused on children's services and on
the criminal justice response to child maltreatment.
Wendy A. Walsh, PhD, is a research assistant professor of
sociology at the Crimes against Children Research Center at
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009
Lippert et al. / Telling Interviewers About Sexual Abuse 113
the University of New Hampshire. Her research includes
studies on evaluating the community response systems to
child abuse, such as the Multi-Site Evaluation of Children's
Advocacy Centers, and research on the criminal justice
response to child victimization. She also has been conducting research on resilience among maltreated children using
the National Survey of Child and Adolescent Well-Being
data; statutory rape victimization; and sexual assault on college campuses.
Lisa M. Jones, PhD, is a research assistant professor of psychology at the University of New Hampshire and research
faculty with the Crimes Against Children Research Center
(CCRC). She has been conducting services and evaluation
research on interventions in response to child abuse for
over 10 years including research on Children’s Advocacy
Center (CAC) programs and services, child abuse trends,
multidisciplinary team collaboration, and media coverage of
child abuse.
Downloaded from http://cmx.sagepub.com at UNIV OF NEW HAMPSHIRE on May 19, 2009